Data Mining and Machine Learning Analysis to Find Polymers for Electronic and Photovoltaics Applications: A Goal to Achieve Higher Dielectric Constant

作者：Bo Xiao, Nafees Ahmad, Asif Mahmood, Mohamed H. Helal · 发表于：Advanced Theory and Simulations · 年份：2025 · DOI：10.1002/adts.202500166 · 被引用次数：5 · 研究领域：Machine Learning in Materials Science、Computational Drug Discovery Methods、Conducting polymers and applications

Abstract The discovery of polymers with high dielectric constants is of significant interest for advanced electronic applications, such as capacitors, flexible electronics, and energy storage devices. In this study, data mining and machine learning (ML) techniques are applied to identify polymers with superior dielectric constant. Molecular descriptors are calculated. These descriptors are used to train several machine learning models, including linear regression, gradient booting regression, histgradient boosting regression, bagging regression, decision tree regression, and random forest regression. By employing cross‐validation and hyperparameter tuning, best model is optimized for robust predictive performance. A database of 10k polymers is generated and their dielectric constant is predicted best ML model. Thirty polymers with higher dielectric constant values are selected. This work demonstrates the power of data‐driven approaches in accelerating the discovery of high‐performance polymers for electronic applications.